Enterprise Setup Guide · A mail server or mail GW is required for sending e-mail notifications. For proper configuration you will need to open ports as described in section 4.2 below

Polarion® Enterprise Setup Guide

Multi-server Installation & Configuration

Polarion Enterprise Setup Guide Version: 2016 SR1 Revision: 823154 - 20160615 Copyright © 2016 Polarion Software Permission is granted to use, copy, and distribute this document provided that the content is not modified and the above copyright notice and this statement appear in all copies. Polarion is a registered trademark of Polarion Software. Polarion ALM, Polarion REQUIREMENTS, Polarion QA and Polarion VARIANTS are trademarks or registered trademarks of Polarion Software.

Acknowledgements Apache and Subversion are trademarks of Apache Software Foundation. Java is a trademark of Oracle Corporation. Microsoft and Microsoft Windows are registered trademarks of Microsoft Corporation. Linux is a trademark of Linus Torvalds. All other trademarks referenced are property of their respective owners. No relationship or endorsement of or by Polarion Software is implied by the appearance of third party trade or service marks in Polarion documentation.

Polarion Enterprise Setup Guide1 Terminology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

2 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

3 Details . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

4 Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

4.1 Server Software Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

4.1.1 Requirements for Windows installation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

4.2 Server Hardware Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

4.2.1 Coordinator Server . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

4.2.2 Stand-alone Instance Server . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

4.2.3 Cluster Instance Server . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

4.2.4 Shared Services Server . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

4.2.5 Example Hardware Configurations for Instance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

4.3 License requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

5 Installation Use Cases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

5.1 Common terms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

5.2 Setting up a Cluster from New Installations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

5.2.1 Configuring the Cluster's Coordinator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

5.2.1.1 License Deployment on Coordinator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

5.2.2 Configuring the Cluster's Shared Services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

5.2.3 Configuring the Cluster's Nodes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

5.2.3.1 Synchronizing Time on Cluster Nodes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

5.2.4 Configuring the Cluster's Activation Application . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

5.3 Setting Up With Multiple Stand-alone Instances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

5.3.1 Configuring the Coordinator for Multiple Stand-alone Instances Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

5.3.2 Configuring Instance1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

5.3.3 Configuring Instance2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

5.3.4 Access URLs for Multiple Stand-alone Instances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

5.4 Migrating From a Pre-2014 "Multi-Instance" Installation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

5.4.1 Configuring the Coordinator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

5.4.2 Migrating a Remote Instance to a Non-clustered Stand-alone Instance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

5.4.2.1 Checking the Migration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

5.4.3 Moving Local Instances for the Multiple Stand-alone Instances Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

5.5 Updating a Multiple Stand-alone Instances or Cluster setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

6 Configuring Shared Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

6.1 Linux Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

6.1.1 NFS Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

6.2 Windows Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

6.2.1 CIFS/Samba Share Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

6.2.1.1 Shared Services Machine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

6.2.1.2 Node Machines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

7 Security Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

7.1 Recommended Security Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

7.2 Advanced Security Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

Polarion Software

http://www.polarion.comPolarion Enterprise Setup Guide (rev. 823154)

1 | Page LiveDoc by: Polarion ALM 2016 2016-06-15 12:00

http://www.polarion.com


7.3 Authentication for Server Monitoring . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

8 Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

9 Troubleshooting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

10 Appendix: Polarion Instance Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

Polarion Software





1 Terminology

Instance - Machine with one running Polarion installation.

Stand-Alone Instance - An Instance, with its own repository on the same machine (i.e. not clustered), exposed to

the user as one Polarion Server.

Cluster - A group of Instances accessing the Shared Services, exposed to the user as a one logical Polarion

Server.

Coordinator - A specially-configured Polarion installation that manages communication and licenses among

Instances.

Shared Services - A machine that hosts the Subversion repository, shared data and load balancer (user entry

point). There is one Shared Services per Cluster.

2 Overview

The following figure shows one clustered setup and several stand-alone Instances, each with its own repository.

3 Details

A cluster setup requires one dedicated physical or virtual machine for each Instance. Coordinator and Shared Servicesmust be dedicated machines as well.

Polarion Software





Coordinator

Distributes tasks that need to be executed on exactly one Instance in the Cluster.

Serves as Licensing Server for all Instances connected to the Coordinator.

Provides a single entry point to all logical Polarion Servers that are connected to the same Coordinator.

Reconfigures the Load Balancer if some of the Instances are offline.

Shared Services

Provides the Load Balancer that forwards users to a specific Instance in the Cluster.

Entry point for one Cluster (on the picture above it is entry point for Cluster).

Provides a File Repository shared by all Instances in the Cluster.

Serves the Subversion repository, which contains the data for the clustered logical Polarion Server.

Instance1, Instance2

Machines running Polarion Service, connected in a Cluster and all configured to use the same Shared Services.

Every instance in the Cluster has its own Polarion data (indexes, object maps), and PostgreSQL database.

(Shared database is not currently supported.)

4 Requirements

Several virtual or physical machines are needed: one for Coordinator, one for every Instance (stand-alone or fromCluster) and one Shared Services per Cluster. The following sections briefly describe the requirements for all theseservers.

4.1 Server Software Requirements

Same for all the machines as described in the Polarion Installation Guides (for Windows and Linux), with one exception:minimum Apache HTTP Server version is 2.2.17 (latest 2.4.x is recommended).

Although the Coordinator machine does not really need Subversion, it is still recommended to use the standard PolarionInstaller to install Polarion on it. It will install Subversion on the Coordinator, and can just remain there.

Coordinator, Nodes, and Stand-alone Instances must all be running the same version of Polarion.

4.1.1 Requirements for Windows installation

A Polarion clustered setup in a Microsoft Windows environment requires the following:

MS Active directory.

A DNS service is recommended, but you can also use static IP addresses (the same subnet is expected).

Testing for proper network configuration using 'ping' between all hosts in a cluster.

A domain user (example: yourDomain\polarion).

For shared Services: CIFS/Samba requires a domain User (example: yourDomain\polarion).

A mail server or mail GW is required for sending e-mail notifications.

For proper configuration you will need to open ports as described in section 4.2 below.

4.2 Server Hardware Requirements

Generally the requirements for all server machines are similar to those described in the Polarion Installation Guides for

Polarion Software





Linux and Windows. All machines must be connected by a fast 1 Gbps low-latency (< 10 ms) intranet network.

4.2.1 Coordinator Server

Requirement Description

CPU 1 - 2

RAM 1GB

Disk Space 10GB

FQDN e.g. coordinator.mycompany.com

Access from Clients (http(s)) Coordinator provides signpost and server monitoring pages. Choose the port (usually

80 for http, 443 for https), configure Apache and Polarion to use it.

Related Polarion properties:

base.url=http(s)://host:port (FQDN or IP address)

Access from Instances Communication between instances and Coordinator (ZooKeeper) takes place on

TCP/IP port specified by property com.polarion.zookeeper.port of the Coordinator, by

default it is 2181. This port on Coordinator host must be accessible by all instances.


com.polarion.zookeeper.port=port# (on Coordinator)

com.polarion.zookeeper=host:port# (on Instances)

4.2.2 Stand-alone Instance Server


CPU 4 - 32

RAM 4GB, recommended 8GB or more

Disk Space 20GB, recommended 250GB or more

FQDN e.g. myserver1.mycompany.com

Access from Clients (http(s)) Choose the port (usually 80 for http, 443 for https), configure Apache and Polarion to use

it.


base.url=http(s)://host:port (Must be FQDN or IP address)

Access to Subversion The same as in the case of a simple installation. There should be http(s) access from

clients (end users), svn protocol access is recommended for fast local access by system

user.


repo=http(s)://host:port/repo

repoSystem=[svn/file/http(s)]://host:port

Polarion Software



https://www.polarion.com/resources/download/polarion_install_guide_linux

https://www.polarion.com/resources/download/polarion_install_guide_win



4.2.3 Cluster Instance Server


CPU 4 - 32



Time Synchronization Well synchronized system time with all other cluster instances.

Access from Load Balancer Load balancer needs to be able to redirect the requests to the cluster instances using on

http(s) port where the Polarion is running.


base.url=http(s)://host:port (Must be FQDN or IP address)

com.polarion.loadBalancer.workerUrl=http://host

Communication between Cluster Instances RPC communication between cluster instances takes place on TCP/IP port specified by

property controlPort of the instances. All instances of the cluster have to be able to

access control ports of all other instances in the cluster.


controlPort=port#

controlHostname=host (Must be FQDN or IP address)

4.2.4 Shared Services Server


CPU 2 - 8



FQDN e.g. myserver2.mycompany.com

Access from Clients to Load Balancer

(http(s))

Entry point to the cluster. Choose http(s) protocol, configure Apache, adjust configuration of

cluster instances.


base.url=http(s)://host:port (on instances - Must be FQDN or IP address)

Access from Coordinator to Load Balancer

Manager (http(s))

Coordinator communicates with load balancer manager via http(s). Configure load balancer

manager application location in Apache.


com.polarion.loadBalancer=http(s)://host/balancer-manager (on cluster instances)

com.polarion.cluster.#ClusterId#.loadBalancer.user= (on coordinator)

com.polarion.cluster.#ClusterId#.loadBalancer.password= (on coordinator)

Shared Folder (Linux paths are used, use analogical paths on Windows.) Folder /opt/polarion of

Shared Services has to be mounted as /opt/polarion/shared on all cluster instances.

This folder sharing should be set up after installation of Polarion. User "polarion" on all

nodes must have read access for /opt/polarion/shared/**, and write access at least

for:

Polarion Software





/opt/polarion/shared/data/svn/*

/opt/polarion/shared/data/BIR/**

/opt/polarion/shared/data/RR/**

/opt/polarion/shared/data/workspace/**

(/* means files inside directory, /** means everything including subdirectories recursively)

Files created by user on behalf of which polarion service is running ("polarion") on any

node in /opt/polarion/shared must be readable by user on behalf of which Apache

server is running on the shared services.


com.polarion.shared=/shared/directory/path (on instances)

Access to Subversion From clients (end users) and each instance of the cluster, the Subversion repository must

be accessible. Either http(s) or svn protocol can be used (svn is recommended for fast

access by the system user).


repo=http(s)://host:port/repo (on instances)

repoSystem=[svn/http(s)]://host:port (on instances)

4.2.5 Example Hardware Configurations for Instance

S M L XL XXL

Operating System 32 or 64-bit 64-bit 64-bit 64-bit 64-bit

CPU Cores 4 8 8 16 32

GB RAM (dedicated to

Polarion)

8 (4) 16 (8) 32 (16) 64 (32) 128 (64)

Storage for Polarion 250GB+ 500GB+ 1TB+ (SCSi or

similar)

1TB+ (RAID 10, NAS,

SAN)

1TB+ (RAID 10, NAS,

SAN)

# Projects < 100 > 100 > 300 500 500

# Users < 50 < 150 > 150 > 300 > 500

Make sure that there is enough RAM available to the OS for file-caching. If SVN is hosted on a different machine, morememory could be allocated for the Polarion process.

4.3 License requirements

If you will host Polarion on your company’s infrastructure, you must provide all physical hardware and/or virtualmachines needed for the setup you want to implement (see Chapter 5 Installation Use Cases), and obtain a license forthe instances you will run. If you utilize Polarion’s cloud-based hosting services, you need to order a virtual server foreach Instance of a clustered or multiple stand-alone instance configuration.

Every node in a cluster or server in a Multiple Stand-alone Instances setup counts towards the “multiple instances” limitset in the license. Please contact Polarion Software for assistance with any licensing questions.

5 Installation Use Cases

This chapter describes simple use cases: a Cluster and a Multiple Stand-alone Instances setup. A subsequent chapterwill describe how to migrate a pre-2014 "multi-instance" setup to the new configuration (released with version 2014). The

Polarion Software



http://www.polarion.com/company/contact/contact.php



"multi-instance" setup with local Instances configured with a pre-2014 of Polarion still works with Polarion 2014 andnewer release without any changes in the configuration. However, it is no longer possible to create new local instances.

If you want to configure a clustered Instance from any of your local Instances, then you need to migrate the whole setupto the new Multiple Stand-alone Instances setup, where Instances always run on a dedicated machine.

5.1 Common terms

[INSTALL] - The root directory of your current installation. This would typically be C:\Polarion on Windows or

/opt/polarion on Linux.

[APACHE] - Apache configuration directory. On Linux should be /etc/httpd/conf.d/ and on WindowsC:\Polarion\bundled\apache\conf\extra

5.2 Setting up a Cluster from New Installations

This section describes how to set up a simple Cluster with two Nodes from new/clean Polarion installations. Thesemachines must be running the same version of Polarion.

Prerequisites - 4 machines (virtual or psychical):

Coordinator (http://coordinator.yourdomain.com)1.

Node1 (http://node1.yourdomain.com)2.

Node2 (http://node2.yourdomain.com)3.

Shared Services (http://cluster.yourdomain.com)4.

Deployment diagram:

Start by installing the same version of Polarion on each of the machines: Coordinator, Node1, Node2 and SharedServices.

Note that different third-party software is required on individual machines:

On the Nodes - Java, Apache HTTP Server, and PostgreSQL

On the Coordinator - Java, and Apache HTTP Server

Polarion Software





On the Shared Services - Apache HTTP Server, and Subversion

The easiest way is to use the standard installation procedure to install all dependencies and eventually uninstall thesoftware that is not needed, if you need to save space on the storage. This software is bundled in Windows distributions,and is already present on most Linux installations. Refer to the Polarion Installation Guide for your preferred operatingsystem for complete installation instructions. (This PDF guide is bundled in all Polarion distributions and is also availablefor download on the Polarion Software web site at: https://www.polarion.com/guides-and-manuals.)

Instantiation of a local Subversion repository must be done only on the Shared Services machine, as it is the onlyrepository that will actually be used. Polarion should not be started immediately after installation, as furtherchanges in configuration are required in order to set up the cluster.

Because Coordinator serves as a license hub for all the nodes and instances connected to it, you do not need to activateany licenses on the nodes.

The next steps in the process assume that Polarion is successfully installed on each node. There are specific next stepsare specific for each machine.

5.2.1 Configuring the Cluster's Coordinator

To configure the Coordinator:

Stop Polarion.1.

Make a backup of the original polarion.properties file.2.

Replace polarion.properties using the Coordinator template file provided in [INSTALL]/polarion/install3.folder: polarion.properties.template.coordinator. (Replace the original polarion.properties file with a copy

of the template file, renaming that as polarion.properties).

Make the following changes in the template-derived properties file, following comments from the template:4.

- Specify base.url appropriately for the machine. Must be FQDN or IP address.

- Set the same value for ajp13-port as in the original polarion.properties file.

- Set the same value for controlPort as in the original polarion.properties file.

- Specify controlHostname appropriately for the machine.

(Optional) Uncomment two properties about load balancer credentials if Apache Load Balancer is protected using5.

basic authentication according to step 3 in chapter 5.2.2 (username and password).

- The default setup uses the same credentials as the svn repository.

(Optional) Change the ZooKeeper port if the default port specified is not appropriate or blocked by firewall policy.6.

(Optional) To disable the unused SVN repositories on the nodes remove the file polarionSVN.conf from the7.

apache configuration directory and restart Apache.

- apache configuration directory on Linux should be: /etc/httpd/conf.d/

- apache configuration directory on Windows should be: C:\Polarion\bundled\apache\conf\extra

(Windows only) You must make sure that the polarion service is started with credentials of a domain user created8.

for polarion - the same user for all polarion installations.

Start Polarion.9.

Polarion Software



https://www.polarion.com/guides-and-manuals



Below is the configured polarion.properties file for Coordinator from the steps above:

com.polarion.application=polarion.coordinatorbase.url=http://coordinator.yourdomain.comTomcatService.ajp13-port=8889 # Control port and host name for shutdown requestscontrolPort=8887controlHostname=coordinator.yourdomain.com # Credentials used to connect to the load balancer, if authentication is enabled# Replace #ClusterId# with the id of the cluster.#com.polarion.cluster.#ClusterId#.loadBalancer.user=#com.polarion.cluster.#ClusterId#.loadBalancer.password= # Port to connect to zookeepercom.polarion.zookeeper.port=2181

5.2.1.1 License Deployment on Coordinator

Polarion 2015 and later: Activate Polarion license using the Polarion Activation screen on the Coordinator.Accessing http://coordinator.yourdomain.com/polarion or login screen of any Node or Instance will redirect youautomatically to the Polarion Activation screen. For more information see Activation Help.

Polarion 2014 and earlier: Make sure that the correct license file is placed in the license folder prior to starting theserver:- on Linux: /opt/polarion/polarion/license/- on Windows: C:\Polarion\polarion\license\

A Cluster's license is activated in the same way as a single instance (described in the Polarion Installation Guidedocumentation). The activation application runs on the Coordinator machine and instructs the user how to activate on-line or off-line. Users accessing any entry point and the login screens of individual nodes and instances are redirected tothe activation page on Coordinator until activation is complete. Nodes and instances can start even if Polarion is notactivated, but users cannot log in.

5.2.2 Configuring the Cluster's Shared Services

Stop Polarion server.1.

Uninstall the Polarion service.2.

- On Linux, run this script: /opt/polarion/bin/uninstall_polarion_service.sh

- On Windows, run: C:\Polarion\polarion\service.bat -uninstall

Configure Load Balancer in Apache using the example template file provided in [INSTALL]/polarion/install3.

folder: loadbalancer.conf.apache24.template (for Apache 2.4) or loadbalancer.conf.apache22.template

(for Apache 2.2). Copy it to [APACHE] directory and rename it to loadbalancer.conf- Basic authentication is configured in the template and you need to check the correct location for theAuthUserFile.(Windows only) Make sure that loadbalancer.conf is included in httpd.conf:4.

# PolarionInclude conf/extra/loadbalancer.confInclude conf/extra/polarion*.conf

Polarion Software



http://www.polarion.com/resources/activation.php



(Windows only) Comment out or remove the following lines from polarion.conf:5.

ProxyPass /polarion ajp://127.0.0.1:8889/polarion timeout=600ProxyPassReverse /polarion ajp://127.0.0.1:8889/polarion

Restart Apache.6.

Make changes in the template file, following the comments provided:7.

- Change path for passwd appropriately for this machine.

On Linux it will be /opt/polarion/data/svn/passwd.

On Windows it will be C:\Polarion\data\svn\passwd

- Adjust BalancerMembers to point to the address of each node.

- Adjust ProxyPassReverse to point to the address of each node

- (Optional) Uncomment logging directives if you want enable logging for Load Balancer.

Set up the shared folder:8.

- On Linux machines, we recommend NFSv4 protocol for sharing

- On Windows machines you can use CIFS/Samba share for sharing. It must be shared for the same domain user

that is used for running all polarion installations in the cluster. The user needs All permissions for the share.

- Data sharing for different Operating systems and protocols is covered in the chapter on Configuration

of Shared Data.

Make backup of the original polarion.properties file on this machine.9.

Modify polarion.properties:10.

- The original properties from the clean installation must be preserved. These properties will be shared between

nodes in the cluster, so everything that is common to nodes should be there.

- Add property: com.polarion.zookeeper=coordinator.yourdomain.com:2181

- Add property: com.polarion.clusterId=cluster1

- Add property: com.polarion.clusterLabel=Main Cluster

- Add property: com.polarion.clusterDescription=Description of Main Cluster

- Add property: com.polarion.loadBalancer=http://cluster.yourdomain.com/balancer-manager

- Modify property: svn.access.file=$[com.polarion.shared]/data/svn/access

- Modify property: svn.passwd.file=$[com.polarion.shared]/data/svn/passwd

- Modify property: polarion.build.default.deploy.repository.url=file://$[com.polarion.shared]/data/shared-maven-repo

- Comment out property: repoSystem- Comment out property: com.polarion.platform.internalPG

Note: this URL must point to the Apache Load Balancer Manager URL. Domain is machine-specific and will be

used as the entry point for this cluster.Note: Property com.polarion.platform.internalPG has to be present in all nodes polarion.properties file

Below is the configured polarion.properties for Cluster - Shared Services (this properties file will be used by eachnode in the cluster).

NOTE: You should not use "file" protocol in the repoSystem property while using a cluster setup for performancereasons. Either comment out, remove the line, or else set up a "svn" server to use with this property. If you still want touse the "file" protocol, you need to point it to the shared repository.

Polarion Software





# Newly added properties to original filecom.polarion.zookeeper=coordinator.yourdomain.com:2181com.polarion.clusterId=cluster1com.polarion.clusterLabel=Main Clustercom.polarion.clusterDescription=Description of Main Clustercom.polarion.loadBalancer=http://cluster.yourdomain.com/balancer-manager # Modified properties#repoSystem=…svn.access.file=$[com.polarion.shared]/data/svn/accesssvn.passwd.file=$[com.polarion.shared]/data/svn/passwdpolarion.build.default.deploy.repository.url=file://$[com.polarion.shared]/data/shared-maven-repo # List of properties from original filerepo=…etc..

5.2.3 Configuring the Cluster's Nodes

IMPORTANT: The following steps must be performed for each Node in the Cluster.

Stop Polarion server.1.


Replace polarion.properties using the example template file provided for nodes in3.

[INSTALL]/polarion/install folder: polarion.properties.template.nodeMake sure that the shared folder is mounted on this machine on the recommended path:4.- shared folder on Linux should be in /opt/polarion/shared- shared folder on Windows is accessed directly as \\<shared_services_host>\PolarionMake changes in template file following the comments provided:5.- Set com.polarion.shared to point to the mounted shared folder (on Linux it should be /opt/polarion/shared,on Windows it should be \\<shared_services_host>\Polarion)- Set the same value for ajp13-port as in the original polarion.properties file.- Set the same value for controlPort as in the original polarion.properties file- Set value for controlHostname to node1.yourdomain.com or node2.yourdomain.com (depending on which nodeyou are configuring).- Set the value for com.polarion.loadBalancer.workerUrl to the specific node in cluster so that the loadbalancer knows the URL of the node.- Set a value in the calc.base.url property to a specific node in the Cluster. It must point to the specific node,otherwise calculation will fail. It is the same as workerUrl.(Eg. calc.base.url=http://node1.yourdomain.com).- Add property com.polarion.platform.internalPG (with value) from the Cluster's shared services whereshould be commented out(Optional) To disable the unused SVN repositories on the nodes remove the file polarionSVN.conf from the6.Apache configuration directory and restart Apache.- Apache configuration directory on Linux should be /etc/httpd/conf.d/- Apache configuration directory on Windows should be C:\Polarion\bundled\apache\conf\extra(Windows only) Make sure that the Polarion service is started using credentials of a domain user created for7.Polarion - the same user should be used for all Polarion instances.Start Polarion.8.

Below is the configured polarion.properties file for Node1. (It will be the same for second or third Nodes, except thatURLs must be changed accordingly.)

Polarion Software





# Shared folder between the machines that make up the cluster# default Linux: com.polarion.shared=/opt/polarion/shared#default Windows: com.polarion.shared=\\<shared_services_host>Polarioncom.polarion.shared=/opt/polarion/sharedcom.polarion.nodeId=node1TomcatService.ajp13-port=8889 #Url of node in load balancercom.polarion.loadBalancer.workerUrl=http://node1.yourdomain.com # Control port and host name for shutdown requestscontrolPort=8887controlHostname=node1.yourdomain.com #Node-specific url of the node.#It is used in calculations to access Polarion via web services#calc.base.url=http://example-nodecalc.base.url= http://node1.yourdomain.com #Postgres database connectioncom.polarion.platform.internalPG=polarion:passwordForDatabase@localhost:5433

You have now configured an entire Cluster from new/clean installations of Polarion.

Your cluster will be accessible on: http://cluster.yourdomain.com

Server Monitoring is accessible on: http://coordinator.yourdomain.com/polarion/monitoring

Apache Load Balancer Manager is accessible on: http://cluster.yourdomain.com/balancer-manager

5.2.3.1 Synchronizing Time on Cluster Nodes

IMPORTANT: Time must be synchronized on each node in the cluster on the OS level by a system administrator. Ideallythis should be automated sync via NTP. If time is not synchronized, users will see different times on each node,scheduled jobs may appear to start off schedule and the Monitor will incorrectly order jobs by time.

5.2.4 Configuring the Cluster's Activation Application

Beginning with version 2015, Polarion includes an activation application that makes it possible to install or update alicense during runtime of the Polarion server, without need to copy the license file manually to the target machine.Access to this application is NOT protected initially by user name and password. For production use it is highlyrecommended to secure access to this application directly in the Apache configuration. It is necessary to perform thisconfiguration only on the coordinator server.

In version 2015 installations there is a template Apache configuration file in the Polarion installation folder:/polarion/polarion/install/polarion.activation.conf.template

To ensure that a user name and password is requested when accessing the activation application(/polarion/activate/online and /polarion/activate/offline), copy this file to the Apache configurationfolder. On Linux usually /etc/httpd/conf.d/. On Windows, usually C:\Polarion\bundled\apache\conf\extra\.

After copying the file, rename it to remove the extension ".template". Then open the file in any text editor and modify itaccording to the instruction comments provided.

The template configuration is prepared for both users file authentication (like Polarion uses for Subversion by default,with user passwords data in a file) and for authentication against an LDAP server.

5.3 Setting Up With Multiple Stand-alone Instances

You can set up the Multiple Stand-alone Instances configuration using the Coordinator for license management.

Three machines (virtual or psychical) are required for this setup:

Polarion Software





Coordinator (http://coordinator.yourdomain.com)1.

Stand-alone Instance1 (http://instance1.yourdomain.com)2.

Stand-alone Instance2 (http://instance2.yourdomain.com)3.

Deployment diagram for Multiple Stand-alone Instances

Start by installing the same version of Polarion on each of the 3 machines: Coordinator, Instance1, and Instance2.

Note that different third-party software is required on individual machines:

On the Instances - Java, Apache HTTP Server, Subversion, and PostgreSQLOn the Coordinator - Java, and Apache HTTP Server

The easiest way is to use the standard installation procedure to install all dependencies and eventually uninstall thesoftware that is not needed, if you need to save space on the storage. This software is bundled in Polarion distributionsfor Windows, and is already present on most Linux installations. Refer to the Polarion Installation Guide documentbundled with Polarion distributions and available on the Polarion Software web site at: https://www.polarion.com/guides-and-manuals.

The next sections assume that Polarion is successfully installed using the standard installation and running on eachmachine. There are specific next steps that need to be performed on each machine.

5.3.1 Configuring the Coordinator for Multiple Stand-alone Instances Setup

Configuration of the Coordinator is exactly same as for the Cluster setup described in chapter 5.2 Setting up a Clusterfrom New Installations, section 5.2.1 Configuring the Cluster's Coordinator. Proceed to configure the Coordinator for thissetup as described there.

The next section on Instances configuration will refer to this Coordinator machine(http://coordinator.yourdomain.com) and assumes that the Coordinator is configured and running.

5.3.2 Configuring Instance1

On the machine hosting the Polarion installation for Instance1:

Stop Polarion.1.

Make backup of the original polarion.properties file.2.


Polarion Software







- All properties in the original file must be preserved.

- Add new property: com.polarion.zookeeper=coordinator.yourdomain.com:2181

- Add new property: com.polarion.clusterId=Cluster1

- Add new property: com.polarion.nodeId=Instance1

- Add new property: com.polarion.clusterLabel=First Server

- Add new property: com.polarion.clusterDescription=Description of first Server

Start Polarion.4.

Below is the configured polarion.properties file for Instance1:

# Newly added properties to original filecom.polarion.zookeeper=coordinator.yourdomain.com:2181# com.polarion.clusterId - is it identificator on coordinator# (instance displays as independent cluster)com.polarion.clusterId=Cluster1com.polarion.nodeId=Instance1com.polarion.clusterLabel=First Servercom.polarion.clusterDescription=Description of first Server # List of properties from original filerepo=…repoSystem=…etc..

5.3.3 Configuring Instance2

On the machine hosting the Polarion installation for Instance2:

Stop Polarion.1.



- All properties from original file must be preserved


- Add new property: com.polarion.clusterId=Cluster2

- Add new property: com.polarion.nodeId=Instance2

- Add new property: com.polarion.clusterLabel=Second Server

- Add new property: com.polarion.clusterDescription=Description of second Server

Start Polarion.4.

Below is the configured polarion.properties file for instance2:

# Newly added properties to original filecom.polarion.zookeeper=coordinator.yourdomain.com:2181# com.polarion.clusterId - is it identificator on coordinator# (instance displays as independent cluster)com.polarion.clusterId=Cluster2com.polarion.nodeId=Instance2com.polarion.clusterLabel=Second Servercom.polarion.clusterDescription=Description of second Server # List of properties from original filerepo=…repoSystem=…etc..

The configuration is quite similar to the Cluster setup. The difference is that there is no Load Balancer or Shared

Polarion Software





Services. Each Instance is autonomous, a "stand-alone" Polarion installation with its own SVN repository. IndividualInstances have nothing to do with other Instances in the Multiple Stand-alone Instances setup. However, users caneasily switch between the Instances by accessing the Entry Point on the Coordinator. You can also monitor availability ofeach Instance using Server Monitoring.

NOTE: The Polarion UI and end-user documentation use the term "Server" when referring to what we term "Instance"for administrators. For example, the UI provides end users the possibility to "Change Server". In Administration terms,that is... work on a different Instance.

5.3.4 Access URLs for Multiple Stand-alone Instances

Entry Point (all Instances): http://coordinator.yourdomain.com/polarion

Server Monitoring: http://coordinator.yourdomain.com/polarion/monitoring

Instance1 direct access: http://instance1.yourdomain.com/polarion

Instance2 direct access: http://instance2.yourdomain.com/polarion

5.4 Migrating From a Pre-2014 "Multi-Instance" Installation

Several Polarion versions prior to version 2014 supported a topography of multiple Polarion Instances which was called"Multi-Instance" setup. Instance clustering was not supported. Although existing customer installations of this setup arestill usable with version 2014 and later, the setup is now considered deprecated in favor of the many improvementsdelivered in version 2014, and it is generally recommended to change over such existing installations. This chapterprovides information and procedures. Customers with a current support and maintenance package may consult PolarionTechnical support for assistance with this migration.

The new Multiple Stand-alone Instances setup differs from old Multi-Instance setup in the following ways:

The so-called Master is replaced by the Coordinator, which manages a license for all Instances.

Local Instances are not compatible with the new Multiple Stand-alone Instances setup. For backward

compatibility, all operations on existing local Instances are still maintained after updating to Polarion 2014 or later.

However, you will not be able to create new Instances. If you have local Instances configured and wish to update

to 2014+ Multiple Stand-alone Instances, these local Instances should be moved to separate machines and then

configured later as part of a Multiple Stand-alone Instances setup (see 5.3.3 Moving Local Instances for the 2014

Multi-Instance Setup).

Each Remote Instance will become a non-clustered Instance connected to the Coordinator.

The Coordinator does not start up the Instances... these must be started individually.

In order to do the migration, you need to update Polarion on the old "Master" and "Remote" Instances to the sameversion. Then you need to modify the configuration files so that they reflect the new configuration properties.

For example: A pre-2014 setup with 1 Master application and 2 remote Instances will become a 2014 Multiple Stand-alone Instances setup with 1 Coordinator and 2 non-clustered Instances, each Instance hosting a stand-aloneinstallation of Polarion, complete with third-party software and repository.

5.4.1 Configuring the Coordinator

To replace the pre-2014 Multi-Instance setup you need to configure the Coordinator. Coordinator is still running on themachine where the Master and Local Instances were running.

Follow the steps described in section 5.2.1 Configuring the Cluster's Coordinator, using also the information from_controller.properties if needed. For example, controlPort and controlHostname can be taken from_controller.properties.

From this point on, it is assumed that you have the Coordinator configured, running and accessible through following

Polarion Software





URL: http://coordinator.yourdomain.com/polarion.

5.4.2 Migrating a Remote Instance to a Non-clustered Stand-alone Instance

The following changes need to be done on each Remote Instance.

Stop the Instance and update it to the latest Polarion version (2014 or later) in the usual way.1.

Make backup of the original polarion.properties file.2.

Modify polarion.properties: 3.

- All properties from original file must be preserved


- Add new property: com.polarion.clusterId=OldRemoteInstanceId

- Add new property: com.polarion.nodeId=OldRemoteInstanceId-node1

- Add new property: com.polarion.clusterLabel=Old Remote Instance Label

- Add new property: com.polarion.clusterDescription=Description of the old remote instance

If you have any properties configured in instanceid.properties, they should be moved into4.

polarion.properties otherwise they will be ignored.

Start Polarion.5.

Below is an example of a polarion.properties file for a migrated remote instance (the instance ID is instance1):

# Newly added properties to original filecom.polarion.zookeeper=coordinator.yourdomain.com:2181com.polarion.clusterId=instance1com.polarion.nodeId=node1com.polarion.clusterLabel=Remote instance - Instance1com.polarion.clusterDescription=Description of the remote instance

# List of properties from original filerepo=…repoSystem=…etc..

5.4.2.1 Checking the Migration

To check that the migration was successful, go to http://coordinator.yourdomain.com/polarion and connect to theInstances.

Entry Point URL: http://coordinator.yourdomain.com/polarion

Server Monitoring URL: http://coordinator.yourdomain.com/polarion/monitoring

Each Instance can be still directly accessed through its URL: e.g. http://instance1.yourdomain.com/polarion

The old configuration files for the pre-2014 Multi-Instance setup (from[polarion_installation]/configuration/multi-instance/*) will become obsolete.

5.4.3 Moving Local Instances for the Multiple Stand-alone Instances Setup

Moving a Local Instance means moving the existing repository and the configuration files to a new Polarion installation.

This step is only required if some of the Instances are configured as Cluster. If no Cluster is needed, the Local Instanceswill still work as they did before in the old Multi-instance setup with the same configuration.

Linux paths:

polarion.properties file: /opt/polarion/etc

Repository folder: /opt/polarion/data/svn OR /opt/polarion/data/multi-instance/instanceId/svn

Polarion Software





Windows paths:

polarion.properties file: C:\Polarion\polarion\configuration

Repository folder: C:\Polarion\data\svn OR C:\Polarion\data\multi-instance\instanceId\svn

To move a Local Instance to a new machine:

Install Polarion on the new machine. Make sure it can be started correctly, then stop it and keep it stopped for the1.

next steps.

In the new installation location: make a backup of the repository folder. (This copy will subsequently be referred to2.

as folder. (This copy will subsequently be referred to as svn_backup)

In the new installation location: make a backup of the polarion.properties file. (This copy will subsequently be3.

referred to as polarion.properties_backup).

Copy the repository folder from the old Instance to the new machine at location:4.

Linux: /opt/polarion/data

Windows: C:\Polarion\polarion\data

Copy the polarion.properties from the old instance to the same location on the new machine (see path5.

references above).

Start Polarion. You should have all the data from the old instance. 6.

After successful startup, you can delete svn_backup and polarion.properties_backup7.

At this point you have a clean installation of the latest Polarion version holding the data and configuration of the oldInstance. You can configure this Instance as part of a Multi-Instance setup following the steps described in 5.2 SettingUp a Clean Multi-Instance Installation.

5.5 Updating a Multiple Stand-alone Instances or Cluster setup

When updating either setup, it is important that you first shut down all the instances and the coordinator. After that,you can use the update distribution to update the machines in the setup (see steps below).

IMPORTANT: It is not possible to run a Cluster node with old version of Polarion and at the same time have it connectedto the updated Coordinator. Be sure you schedule extended downtime - enough to install the update on all machinesand for the necessary reindexing.

Update steps:

Stop the Polarion service on all nodes, and finally on the Coordinator.1.

Install the update on the coordinator machine and start Polarion in reindex mode.2.

Install the update and start each of the cluster nodes in reindex mode. IMPORTANT: update the Cluster nodes3.in sequence, so that only one of the nodes does an update of the SVN repository. (Updating the shared servicesserver is not required.)

6 Configuring Shared Data

This chapter provides steps how to configure shared data on Linux and Windows machines. The setup is different foreach platform and all differences are covered in this chapter.

Prerequisites (4 machines, all on same domain):

Polarion Software



https://www.polarion.com/downloads/update



Coordinator (coordinator.yourdomain.com)1.

Node1 (node1.yourdomain.com)2.

Node2 (node2.yourdomain.com)3.

Shared Services (cluster.yourdomain.com)4.

The shared folder has same structure as standard Polarion installation folder, so it is possible to use a Polarion installerto create it:

Install Polarion.1.

Uninstall Polarion service and delete folders that are not needed. Only two folders in the Polarion installation are2.

needed for shared data:

- Linux: /opt/polarion/etc and /opt/polarion/data

- Windows: C:/Polarion/polarion/configuration and C:/Polarion/data

Note: Deleting of other (unnecessary) folders is optional. You can optionally leave the installation folder as it is afterinstallation.

The root of the Shared Services is the polarion folder.

6.1 Linux Configuration

Share the folders among the nodes NFSv4 protocol. Other protocols (such as SSHFS or NFSv3) have known problems,so they must not be used.

6.1.1 NFS Configuration

This section describes an example how to set up folder sharing using the NFS protocol.

To configure folder sharing using NFS:

Connect to the Shared Services machine (http://cluster.yourdomain.com)1.

Edit file /etc/exports and add the following lines:2.

- Add line: /opt/polarion node1(rw,sync,no_root_squash,no_subtree_check)- Add line: /opt/polarion node2(rw,sync,no_root_squash,no_subtree_check) On the Node machines create a folder /opt/polarion/shared3.On the Node machines add following line to /etc/fstab4.- Add line: cluster.yourdomain.com:/opt/polarion /opt/polarion/shared nfs defaults 0 0On all machines run following command:5.# /etc/init.d/rpcbind start# /etc/init.d/nfs start On the Shared Services machine run following command:6.# exportfs -a

And on the node machines mount the shared directory with the command:7.# mount -v cluster.yourdomain.com:/opt/polarion /opt/polarion/shared/Check that shared folder appears on each node in thein the /opt/polarion folder, and make sure that polarion.properties8.on each node points to this location: /opt/polarion/sharedMake sure that each node has rw permissions for the /opt/polarion/shared folder, and all nodes create folders and files with9.the same permissions.

6.2 Windows Configuration

This section describes how to make a shared folder on a Windows machine. We recommend using standard Windowssharing on this platform.

Polarion Software





6.2.1 CIFS/Samba Share Configuration

You need to configure sharing on the Shared Services machine, and all Node machines, starting with the SharedServices machine.

6.2.1.1 Shared Services Machine

A simple example on how to create the shared folder using CIFS/Samba.

To create a shared folder on the Shared Services machine:

Connect to the Shared Services machine (http://cluster.yourdomain.com)1.

Open File Explorer2.

Right-click on the C:/Polarion folder3.

Select Properties.4.

Select tab Sharing.5.

Click the Share... button.6.

Set shared user as the same domain user for all polarion installations in the cluster. The user needs full7.

permissions to the folder.

After you have done with sharing options, click on Share, and then on Done.8.

6.2.1.2 Node Machines

To connect shared folder on each Node:

Connect to each Node (http://node1.yourdomain.com and http://node2.yourdomain.com)1.

Open File Explorer2.

In the left panel, right-click on Computer.3.

Map a network drive. Using the credentials of the polarion domain user.4.

- Folder should be: \\cluster.yourdomain.com\polarion

Edit the polarion.properties file accordingly and specify the path to the shared folder.5.

- Property com.polarion.shared must point to this mapped drive.

7 Security Options

The recommended setup is to use encrypted communication between the outside world and the internal network withservers of the Multiple Stand-alone instances setup (as shown in the figure below). This is also optimal from aperformance point of view. Communication inside the local network can optionally be encrypted as well (except for thefolders shared using NFS). More details about this are provided in the section Advanced Security Options.

Polarion Software





In the sections below it will be noted that https access should be set up in an Apache server. The details about how toactually configure this are out of scope of this guide, however you can see some hints and references in Polarion Help.

7.1 Recommended Security Options

Service Security Settings

Entry point

(Coordinator)

The entry point, where users can select Polarion server, should be configured for https access in Apache, so

that end users will access e.g. https://coordinator.mycompany.com/polarion.

Additional steps:

Remember to update base.url in Polarion properties.

Server monitoring

(Coordinator)

The same as above for the server monitoring page, e.g.

https://coordinator.mycompany.com/polarion/monitoring. This will be usually done by the same configuration as

the entry point.

Stand-alone Instance

Polarion

Standard https setup as for simple stand-alone installation, so that the instance can be accessed as e.g.

https://instance1.mycompany.com/polarion.

If the Subversion repository is accessed by end users, it should be configured for https access as well.

Polarion Software





Additional steps:

Remember to update base.url in Polarion properties.

Clustered Instance

Polarion

Standard https setup in Apache for the Load Balancer, so that the clustered instance can be accessed as

e.g. https://instance2.mycompany.com/polarion. If the Subversion repository is accessed by end users, it should

be configured for https access as well.

Additional steps:

1) Note that in this case it is necessary to set Polarion property wikiProtocolSchema=https in the shared cluster

properties (/opt/polarion/etc/polarion.properties on the Shared Services).

2) Remember to update base.url in shared cluster properties.

7.2 Advanced Security Options

If desired the internal network communication among the servers comprising the Multiple Stand-alone Instancessetup can be encrypted as well.

Service Security Settings

Load Balancing Load balancing communication between load balancer and the workers (clustered instances) can be

done via https if desired. It is necessary to set up https access on Coordinator and all cluster Instances

(as for simple installation), and then configure the load balancer to use the https worker URLs. You can

use the same wildcard certificate on all servers.

Additional steps:

1) It is necessary to switch on SSL proxy engine using SSLProxyEngine on in the Apache

configuration.

2) It is necessary to set Polarion property wikiProtocolSchema=https in the shared cluster properties

(/opt/polarion/etc/polarion.properties on the Shared Services).

3) Remember to update base.url in shared cluster properties.

Note that by default, Apache does not verify the certificate of the workers. To switch it on, set

SSLProxyVerify=require and you might also need to set SSLProxyCACertificatePath or other

directives. See mod_ssl documentation (http://httpd.apache.org/docs/2.2/mod/mod_ssl.html) for more

details.

Load Balancer Management By default, the Coordinator manages the load balancer. For example, it switches off the worker if a

Polarion cluster node disconnects from the cluster. This management is done using the http/https URL

provided by the shared cluster property com.polarion.loadBalancer. The load balancer manager is a

web application provided by Apache, and it can be configured for https access (on Shared Services).

Additional steps:

1) Remember to update Polarion property com.polarion.loadBalancer in shared cluster properties.

2) It might be necessary to install a trusted certificate authority to Java trust store, especially if a self-

signed certificate is used.

Subversion Repository For best performance, cluster nodes should access the shared repository by the system user using the

SVN protocol (repoSystem=svn://...). This makes it necessary to open remote access

to svnserve running on the Shared Services machine. This communication is not encrypted.

To enhance security you may want to consider establishing a secure channel using, for example,

stunnel (https://www.stunnel.org/). The idea is: instead of...

repoSystem=svn://SHARED_SERVICES_HOST/opt/polarion/data/svn/repo

use...

repoSystem=svn://localhost/opt/polarion/data/svn/repo

Polarion Software



http://httpd.apache.org/docs/2.2/mod/mod_ssl.html

https://www.stunnel.org/



...on the cluster node and connect localhost:3690 to SHARED_SERVICES_HOST:3690 by a secured

channel.

Cluster Coordination Instances and Nodes in the Multiple Stand-alone Instances setup communicate with the Coordinator

machine. This communication is not encrypted. It can be secured using, for example, stunnel in a very

similar way to that described above. Cluster instances also communicate directly with other cluster

instances using TCP socket (on the controPort). This communication is encrypted internally.

7.3 Authentication for Server Monitoring

After initial installation of a cluster, the Server Monitoring page is available on the Coordinator. On this page,administrators can view and access all the configured nodes (servers), and manage the load balancer. The access URLfor this page has this form: http://coordinator.yourdomain.com/polarion/monitoring.

The page does not require any authentication. However, authentication is recommended and you can configure basicauthentication via the standard way in Apache, using one of the following directives (assuming the password file iseither /opt/polarion/etc/passwd or C:/Polarion/data/svn/passwd - use the one applicable for your OS):

Apache 2.4 and newer:<Location /polarion/monitoring> Require all denied AuthType Basic AuthName "Monitoring" AuthUserFile "C:/Polarion/data/svn/passwd" Require valid-user</Location>

Apache 2.2 and older:<Location /polarion/monitoring> Order Deny,Allow Deny from all AuthType Basic AuthName "Monitoring" AuthUserFile "/opt/polarion/etc/passwd" Satisfy Any require valid-user</Location>

8 Notes

Extensions are not shared among the Nodes in a Cluster. Each node has its own independent extensions folder

(e.g. /opt/polarion/polarion/extensions). Therefore, an extension can be installed on some specific Node(s)

in a cluster. However, in most cases you will want to install extensions on all Nodes. Only in the case of some

special kind of extension would you not install it on all Nodes.

Scheduled jobs should be reviewed for a cluster, and convenient Node selectors (i.e. the node attribute of<job>

elements) should be specified depending on the nature of the job. The following default jobs should have

node="*" specified: Index Checker, Suspend DB History Creator, Resume DB History Creator.

Diagnostics: Polarion comes with a self-diagnostic utility Polarion Diagnostic Tool (PDT), which can run

comprehensive diagnostic tests and communicate the results to Polarion's technical support team. PDT checks if

Polarion is running in a

cluster and gathers configurations from shared folders. The utility is located in the diagtool folder under the root of

any installed Polarion instance, which also contains documentation for its use.

Polarion Software





9 Troubleshooting

(Linux) It is recommended to disable SELinux, if it is used.

(Windows) Disabling firewall on enterprise editions of Windows also disables crucial network services.

After encountering problems with activities (e.g. org.apache.lucene.index.IndexNotFoundException: no segments*file found in MMapDirectory@/opt/polarion/shared/data/workspace/polarion-data/index/Activities), it is necessaryto manually delete index of activities from the Shared Folder, by default:/opt/polarion/shared/data/workspace/polarion-data/index/Activities and restart one of the nodes (so that newempty index is created).

10 Appendix: Polarion Instance Architecture

Polarion Software





Documents

Enterprise Setup Guide · A mail server or mail GW is required for sending e-mail notifications. For proper configuration you will need to open ports as described in section 4.2 below